It sounds like something out of a science fiction movie, but unlike some of the technology used in sci-fi blockbusters—like the flying cars from Back To The Future 2—voice search is not something we can only expect to have in the next few centuries.
In 2014, Google commissioned a study by Northstar Research to look into the rising popularity of voice search technology. 1400 American citizens across a range of age groups were asked to participate in a survey, which showed that teens—as always—were ahead of the curve in taking advantage of the then-new innovation.
Not only is the function more popular with the younger generation (55% of teens as opposed to 41% of adults), they’ve also explored new ways to integrate voice search in their everyday lives.
Google Zurich’s Principal Engineer, Behshad Behzadi, made a keynote speech at SMX West in March 2016 on the prevalence of voice search in many people’s everyday lives and referenced the Northstar Research study.
Millennials have been found to use voice search most often to start phone calls, followed closely by asking for directions and doing homework research. Other traditional uses include playing music, looking up movie schedules, and checking the time.
Adults, on the other hand, seem to use voice search mostly for getting directions, dictating texts, and logging phone calls, with far less emphasis placed on time, music, and cinema times.
More recently though, a lot of adults have admitted to using voice search in the kitchen—whether it's to convert measurements or look up a recipe—citing it as the best way to avoid a greasy, powdered, or marinated screen while cooking.
And therein we find the number one allure behind using the voice chat feature—it’s faster, hands-free, and lets you multitask. In fact, the vast majority of participants from the survey said voice search made them more efficient and felt safer too.
When the feature was first released, the word error rate was over 20%, but thanks to constant work on improving the software it’s been down to 8%—which is enormous and impressive progress over such a short space of time.
Behzadi certainly agrees with 87% of the 1400 participants in the 2014 study, testifying that using voice apps such as Google Now, Siri, and Cortana is indeed the future.
Google is certainly taking the helm in this regard by working on developing voice recognition technology that goes a step further: voice understanding. The goal is to optimize the search engine so that the meaning and context behind the words we use is understood.
Most people speak and type differently—especially when using a search engine, adopting a choppier shorthand writing style. Perhaps without realizing it, almost everyone is using keywords rather than natural phrases: for example, “weather New York.” When asking out loud, you’d be far more likely to ask the question in full: “What’s the weather like in New York?”
It means we’ll naturally lean toward asking the whole question rather than blurting out keywords when doing a voice search query, so it’s become increasingly important for search engines to be able to interpret natural speech.
In an attempt to fill this niche, Google has been feeding their AI engine thousands of romance novels to teach it more about human speech patterns, quirks, and idiosyncrasies. And just in case you were wondering why they chose romance novels, it’s because they tend to use a very casual tone, incorporating a wide vocabulary range to convey various ideas, the rationale being that this will help the Google AI learn faster.
It’s already helped the Google search engine improve its understanding of simple questions, and the AI is now capable of responding to basic follow up questions too. Long-term goals may see Google being able to communicate with users as if two humans were having a conversation, but at the moment that’s still the stuff of science fiction.
Google is pushing the smart search frontier in many other areas too. One of these ways is by allowing for spelling corrections in your voice search queries. Something we can all expect is for the search engine to bring up photos of whales when asking for a picture of Wales. You won’t have to do a second search specifying “I want a picture of the country Wales” though—simply clarify by spelling the word out, and Google will correct itself.
This ties-in with the second point—search memory. With Google’s smart voice search implementations, you’ll be able to ask follow-up questions with ease. For example, if you ask “Where is Mount Rushmore?” and then “Who are the faces?” Google will understand that you mean US Presidents George Washington, Thomas Jefferson, Theodore Roosevelt, and Abraham Lincoln.
A further extension is Google voice search’s ability to provide results based on what you see on-screen at the time of your query. Reading a Wikipedia entry on John Cusack and want to get a full list of his movies, or see more photos? Just ask for “photos of John” or “movies John is in.” Google will automatically recognize that Cusack is the John you’re looking for, and will direct you to his IMDb profile or Google Images results.
Alternatively, you can also use this feature to help Google narrow down the answer with context if the search engine is unable to respond to a query immediately. A typical example might be “How high is Rigi?” A voice search may return or give information on the wrong Rigi (a meme with someone on drugs, for example). This outcome is because Rigi is an incredibly obscure mountain in the Swiss Alps. Asking “What are the mountains in Switzerland?” first and following up with “How high is Rigi?” gives Google context, allowing the search engine to tell you that Rigi is 1,798m high.
If you need to know how to get from point A to point B or even just the distance, Google voice search will now give you the relevant information based on your location. For example, if you’re at 34th Street and need to get to the Statue of Liberty, you can ask “How far is it to the Statue of Liberty?” and Google will understand that you mean from where you are on 34th Street. Or if you ask “train route to Boston,” you’ll receive directions to the nearest subway station and details on the platforms and trains you’ll want to take.
Even more exciting is Google voice search’s ability to understand app-based context. For example, say you’ve been chatting with a friend on an app like Viber and are trying to decide on a restaurant. If one of you have mentioned a particular restaurant and you want to see their menu, you can simply ask Google to “show me the menu.” Using your chat as a reference for context, Google will immediately know which restaurant menu to bring up for you.
We may have saved the best for last. Because of our busy lifestyle, we don’t often get a chance to go to the shopping mall or retail stores anymore. And when we do, it’s usually absolute chaos because of the sheer number of customers. As a result, more and more people are doing their shopping online.
In the same SMX West keynote speech we mentioned earlier, Behzadi also revealed that Google was working on developing an extension to regular voice search, with online shopping as the focus. He dubs it conversational shopping—and we love it.
But what is conversational shopping exactly? Well, it applies the same principles to online shopping as regular voice search queries. For example, if you’ve done a search for blue jeans, you can then narrow your search by asking for a particular price range, size, or brand.
If you know what you want, you can even tell Google to “order me a pair of size 12 bootleg jeans from American Eagle”. This command creates an instant purchase, and will even use your location as a delivery point!
You won’t be confined to online shopping of course. If you do want to make a trip to the mall but aren’t interested in fighting the crowds while searching for a product that matches your needs, you can also ask Google “what shops are around me?” In using your location and recent search as context, the search engine will then give you a list of available options.
Don’t feel that you’d be limited to clothing, gadgets, and other paraphernalia either. Had a long day at the office and looking forward to last night’s leftovers, only to find that your kid or housemate already had it for lunch? Make an instant purchase by asking voice search to “order me a large chicken supreme pizza from Pizza Hut.”
While voice search has been around (and growing in popularity) since 2014, it’s now at the point where optimizing your site for voice search results is of crucial importance.
Voice search is, after all, third in a list of the largest SEO trends, according to a survey of 39 SEO experts. It follows after mobile optimization, which does, in fact, include voice search optimization, and quality content.
In another I/O keynote speech, with the introduction of Google home (an Amazon Echo competitive feature), it was revealed that in 2016 as much as 20 percent of all search queries in the US were made using Google’s Android App voice search function. This information poses an interesting conundrum for SEO.
Because voice search queries are so different from typed commands (we spoke earlier about the language used), it naturally follows that how site owners and managers approach SEO will have to change somewhat.
Now, it’s understandably difficult to figure out what one needs to focus on when looking forward concerning SEO. After all, there are thousands of articles written each year claiming that SEO has changed drastically from one year to the next, and each has their take on how best to adapt. But despite the number of assertions, the factors determining SEO ranking are unlikely to change much from year to year.
Technological innovations like voice search, on the other hand, do tend to require some significant adaptation in a relatively short space of time. Once again, this has a lot to do with the way users are formatting their search queries regarding language.
Many SEO experts have continuously promoted the use of short keywords, also known as head terms. These typically range from one to three words. While not allowing for much specificity, they’re still incredibly prevalent in textual search queries.
“Eggs” is a very broad short keyword that probably won’t yield the desired search results. As a result, it won’t make much impact on your site’s keyword ranking either. “Scrambled eggs.” on the other hand, is a short keyword phrase that is far more likely to boost your SEO ranking somewhat and provide more detailed results for users. “Scrambled eggs recipe” even more so!
But let’s be honest—regarding voice search queries, short keywords are less likely to boost your site because people don’t speak that way. As mentioned earlier in this article, a voice search query is more likely to look something like this: “how to make scrambled eggs?”
Voice search queries aren’t short keyword based; they’re long keyword and question-based. This notion means the way you use keywords has to change.
So, long tail keywords, which are typically made of three to five words (or more!), are better targeted to specificity. They give the search engine a better idea of what the user is looking for than head terms do. More importantly, they fall into the category of natural speech far better, meaning they’re more likely to come up in a voice search query.
Some experts claim that keywords are virtually dead thanks to voice search. But while saying keywords are dead makes for a catchy title, it’s simply not true.
Despite Google’s leaps and bounds in developing its AI to understand natural speech better, we as humans still use keywords to learn and form questions and answers. We don’t realize it so much because we don’t think of it as keywords, but it’s true nevertheless. Keywords form a major part of sentence structure, no matter what language you’re speaking. Instead of “keyword,” we call it “subject.”
In the same way, search engines use specific keywords to understand what your page’s content is referring to. Matching keywords to a search query is how we gain traffic.
For this reason, saying intentions and topics are replacing specific keywords would be a truer statement. You can think of it regarding long keywords if it makes it easier for you to understand, but there is an essential difference.
Topics and intention do incorporate long tail keywords and keyword phrasing but aren’t limited to them. Keywords are typically precise, whereas topic and intention based searches—voice search queries especially—will rely somewhat on similar words and phrasing.
This cause is why Google has been feeding its AI romance novels, remember? By devouring natural speech patterns, the search engine is learning how to interpret voice search queries based on topic and intent. Once this is understood, results results are put together by translating the topic and intent into keywords. The major difference is that instead of using a limited scope of keywords according to the exact query submitted, voice search results will be considering a range of similar keywords.
Sound complicated, right? It is.
The long and the short of it is that you’ll still be using short keywords in SEO, but you’ll also have to start focusing on long keywords and similar keywords so that the search engines have more to work with when deciphering your content.
This perception becomes even clearer when you consider the fact that while voice search is becoming increasingly popular, it hasn’t replaced typed searches entirely. In fact, it isn’t likely to do so for some time, despite the increase in voice search technology geared for desktop and laptop devices.
At present, voice search is more or less confined to mobile devices. Again, mobile searches are increasingly popular, but they aren’t a clear-cut replacement for desktop technology long-term.
Brand specific queries can pose an interesting issue for voice search results and SEO, because of the spelling and pronunciation issue.
John Oliver, a British comedian, illustrated this point rather brilliantly, which is why we’ve decided to call it the John Oliver Factor.
Think of companies such as TransUnion, Equifax, and Experian. John Oliver and his team targeted them specifically in illustrating spelling and pronunciation as a potential SEO nightmare by creating parody sites for these brands based on pronunciation. TramsOnion, Equifacks, and ExperiAnne are likely to show up on your voice search results pages, which is a major issue for the real companies. It becomes even more alarming when you realize that these are credit brands.
While it’s all good and well that the famous comedian’s parody sites show up to prove a point, hackers have likely explored this option as well. This route makes voice search a bit of liability unless you’re 100 percent certain that the site you’re visiting is the real thing.
In fact, in the research Oliver and his team conducted in preparation for the parody skit they discovered that as many as 10 million US citizens had already fallen prey to this vulnerability. As a result of mistaken identity, they now have significant errors on their credit reports—and the credit reporting companies don’t have a system in place to help fix those blunders.
Thanks to Google allowing for spelling corrections in voice search (see above), consumers do have some ability to find the right site. But this does require that they know how to spell your brand name correctly, which isn’t something you should rely on.
So what do you as a brand company do to prevent users from struggling to find your site or products, or falling prey to malicious parody sites engaging in fraud or selling imitation products?
Firstly, you’ll want to research misspellings that are likely to result from common mispronunciations of your brand name. Getting a group of participants with various accents and dialects together to search for your company using voice search can be a great way to do so. These misspellings and mispronunciations should add to your keyword optimization strategy as a test. Be careful to add relevant negative keywords too, though.
Keywords have never been the only means of improving your SEO rating. With the increased complications of developing a good search engine ranking due to voice search, the other options have become even more prevalent.
One of these options is to implement and improve your URL structure using hierarchy. This structure mainly helps you to build out sections of your website based on topic, which is a major factor in voice search queries’ natural language.
The categorization and layout of your site have a lot to do with this, and you would be well-advised to look into performing usability evaluations and card sorting sessions to find the optimal hierarchy for your site.
You should also be focusing on the type of results voice search tools are likely to give for general searches in your industry. For example, if you ask “find a LASIK eye surgeon near me,” your voice search engine is likely to provide you with Yelp listings for LASIK eye surgeons in the area.
As a business owner, and especially as a site owner or manager, this means you need to put effort into improving your Yelp listing. This effort, in turn, optimizes your chances of appearing in a voice search query on a results page.
Page speed has always been a major consideration in search engine optimization. With the rise of voice search technology, it’s become an even greater ranking factor.
Anything and everything you can do to increase your page speed must be taken into consideration. Condensing your content size (particularly images, videos, and other media) without compromising on quality is one of the simplest means of doing so. To do it right, you’ll need to find out what plugins are available for your site builder and content management system.
Speaking of images, these will be a major step in optimizing your site for voice search results. You should always ensure that your company has good image content, but it’s becoming even more important that your imagery is enticing. It should stand out from your competitors and bring you to the fore among the clutter and noise online.
With creative imagery, you would be more likely to see the benefits of voice search results, especially regarding conversational shopping. Product listing ads will give you a significant advantage, and ought to be a key factor in your digital ad campaigns as a result.
Paid and organic content alike will have to be developed in a way that answers to conversational voice search queries are meaningful. This concept entails a shift in the content strategy known as tagging.
Tagging focuses on increasingly narrowed niches—which is a major factor in voice searches, as we’ve illustrated throughout this article. Latent semantic indicators are part of tagging strategies, which allows for similar results. There’s already been a significant step in this direction, which will become even more important as conversational shopping becomes even more prevalent.
Schema markup is an HTML add-on designed to help search engines get acquainted with your site content’s context. It goes a long way toward ranking in specific queries made through voice search.
Because voice search queries are more likely to take the form of questions, you’ll want to research frequently asked questions related to your site and content. Adding an FAQ page incorporating these questions will help your site to rank better in voice search result pages. It helps if you’re able to answer them in a conversational, natural manner.
Voice search is still a relatively alien concept to many of us, who grew up in an age where technology wasn’t able to “talk back.” But as the younger generations have been more exposed to (and are more embracing of) such innovation, their growing online presence and consumerism make voice search not only the future but the present.
It would be disastrous to ignore the prevalence of voice search in modern culture, especially from an SEO view. Hopefully, this article has gone a long way in helping you understand this next step in online search methods, and shown you how to adapt your online platform accordingly.
Create, edit, customize, and share visual sitemaps integrated with Google Analytics for easy discovery, planning, and collaboration.